Add RSVP speech scoring and failure safeguards

[sajjad-mazaheri]

Summary

• Add RSVP speech transcript scoring and target position alignment.
• Normalize multilingual speech responses before scoring.
• Mark technical speech failures as invalid instead of incorrect.
• Preserve existing matrix and experimenter-scoring modes.

Testing

• TypeScript check passed.
• RSVP speech scoring, normalization, alignment, runtime, and validity tests passed.

[Copilot]

Pull request overview

This PR adds position-aware scoring for RSVP automatic speech responses, including multilingual transcript normalization and token-to-target alignment, and introduces safeguards so technical speech failures are treated as invalid (not incorrectly scored) while preserving existing non-speech modes.

Changes:

• Add Unicode- and language-profile-based speech transcript normalization (English/Persian/Arabic) and deterministic token alignment.
• Implement RSVP speech scoring with policy controls (word order, repetition, self-correction) and rich diagnostics.
• Add runtime lifecycle safeguards (cleanup on skip/failure, provider language mapping from condition language) and trial validity gating for QUEST consumption.

Reviewed changes

Copilot reviewed 15 out of 15 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
threshold.js	Updates RSVP-reading block instructions and records/acts on RSVP speech preparation success.
components/trialRoutines.js	Uses RSVP speech validity rules to decide QUEST validity and whether to consume target word sets.
components/rsvpReading.js	Closes active RSVP speech resources when skipping due to bad tracking.
components/instructions.js	Adds instruction text helpers for automatic speech RSVP reading.
components/rsvpSpeech/rsvpSpeechRuntime.ts	Renames language mapping to condition language; adds close-tracking to avoid controller lifecycle races; improves cleanup on failures.
components/rsvpSpeech/rsvpSpeechValidity.ts	Introduces explicit validity/consumption rules for automatic speech trials (pre/post exposure failures, missing registrations).
components/rsvpSpeech/rsvpSpeechScoring.ts	Adds RSVP speech scoring pipeline (normalize → align → classify → score) with diagnostics and policy enforcement.
components/speech/textNormalization.ts	Implements language profiles and Unicode-aware normalization/tokenization with join-control handling.
components/speech/tokenAlignment.ts	Implements monotonic, position-aware alignment with special handling for Persian join/split patterns and diagnostics.
tests/tokenAlignment.test.ts	Adds alignment behavior tests (ordering, duplicates, Persian join/split, merged tokens).
tests/textNormalization.test.ts	Adds normalization/tokenization tests across English/Persian/Arabic scripts and punctuation cases.
tests/rsvpSpeechScoring.test.ts	Adds scoring/policy tests and multilingual edge cases, including Persian-specific behavior.
tests/rsvpSpeechValidity.test.ts	Adds validity matrix tests for runtime failure states vs QUEST consumption.
tests/rsvpSpeechRuntime.test.ts	Updates runtime tests for renamed language resolver and adds lifecycle/cleanup tests.
tests/lifetime.test.ts	Updates global mocks to include RSVP speech runtime state.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.